Page MenuHomePhabricator
Search Advanced Search
    • Diffusion Commit
    • ·Closed
    #lcaaaaaaaaaaa
    • Diffusion Commit
    • ·Closed
    Final NoteDb migration updates Patch-set: 1 Topic: simplify-listToText
    • Diffusion Commit
    • ·Closed
    Create change Uploaded patch set 1. Patch-set: 1 Change-id: I3efcbba1b9adc9a9e32e366e355cb742376cd91b Subject: Simplify mw.text.listToText Branch: refs/heads/master Commit: 1d13fd503a8f593dec339eacf2378b88a23ed653
    • Diffusion Commit
    Final NoteDb migration updates Patch-set: 1 Topic: comboboxgetter
    • Diffusion Commit
    Create change Uploaded patch set 1. Patch-set: 1 Change-id: I306f12cf86c1285edd36999f5bd95595cd7ee08d Subject: Add a getter method for ComboBoxWidget text inputs Branch: refs/heads/master Commit: 6674bd642ca28573d8e0125cf088df8b563ac664
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: -Code-Review I've rebased change If9cf9a now - if that change is merged then this one is redundant. Patch-set: 1 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Code-Review=0
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: I hosed my vagrant installation, which is why I never got round to doing it, but I'll try and get that done today or tomorrow. Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: (That's change If9cf9a52bd4b1bb42cc7f9f1f1096828710cbc52.) Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: Code-Review+1 Looks fine to me. Note that this is pretty much what Gerrit 145547 does as well. Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Final NoteDb migration updates Patch-set: 8 Topic: html-error-rebase
    • Diffusion Commit
    • ·Closed
    Create patch set 7 Uploaded patch set 7. Patch-set: 7 Subject: Improve error messages in mw.html Commit: 77fb57e2172616c1fca89866bac2924b65491ff1
    • Diffusion Commit
    • ·Closed
    Update patch set 6 Patch Set 6: (1 comment) Patch-set: 6
    • Diffusion Commit
    • ·Closed
    Update patch set 6 Patch Set 6: (3 comments) Patch-set: 6 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 6 Uploaded patch set 6. Patch-set: 6 Subject: Improve error messages in mw.html Commit: 51e0e11c39c981dc49cee6bad1a85e02dccd95d3
    • Diffusion Commit
    • ·Closed
    Update patch set 5 Patch Set 5: Now rebased - my first one ever. I inadvertently changed the topic name, but other than that I think it went ok. (And now I know how to do it properly for next time.) Patch-set: 5 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 5 Uploaded patch set 5. Patch-set: 5 Subject: Improve error messages in mw.html Commit: 1142a59ead0133ef340937199738b8e430475589
    • Diffusion Commit
    • ·Closed
    Update patch set 4 Patch Set 4: Patch set 4 makes sure that libraryUtil.checkTypeMulti doesn't alter the expectTypes table that it was passed. Patch-set: 4 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 4 Uploaded patch set 4. Patch-set: 4 Subject: Improve error messages in mw.html Commit: 402e9065ab37793d748739367cfdc847c31f56aa
    • Diffusion Commit
    • ·Closed
    Update patch set 3 Patch Set 3: (1 comment) Patch-set: 3
    • Diffusion Commit
    • ·Closed
    Update patch set 3 Patch Set 3: (1 comment) Patch-set: 3 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 3 Uploaded patch set 3. Patch-set: 3 Subject: Improve error messages in mw.html Commit: 66d5cd2581307282deddaccac36f2ad24431cd4d
    • Diffusion Commit
    • ·Closed
    Create patch set 2 Uploaded patch set 2. Patch-set: 2 Subject: Improve error messages in mw.html Commit: 234583b252d6d70945616471367237e5ae7bef2b
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: (1 comment) Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: (1 comment) Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: (2 comments) Patch-set: 1 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create change Uploaded patch set 1. Patch-set: 1 Change-id: If9cf9a52bd4b1bb42cc7f9f1f1096828710cbc52 Subject: Improve error messages in mw.html Branch: refs/heads/master Commit: 63d0d9086029300cfc38da398dcb450e60bb9768
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: If I'm reading the PHP code right, then this would change the point that errors related to database accesses were raised. At the moment such errors are raised when the mw.title object is created, but it looks like this change would mean they would be raised when one of the expensive properties or methods was accessed instead. I think this is a very good idea, and it will solve quite a few problems I have encountered in writing modules. However, we should also regard it as a breaking change, and advertise it as such. I can think of several widely-used modules that catch errors when creating mw.title objects, but don't do anything special when accessing e.g. title.exists. On pages that exceed the expensive function count limit or have other errors, this would cause those modules to give script errors where they weren't previously. A couple of examples of affected modules: https://en.wikipedia.org/wiki/Module:Redirect https://en.wikipedia.org/wiki/Module:Archive_list (My PHP skills are still quite basic, so I'm not going to give this a +1 or -1.) Patch-set: 1 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Final NoteDb migration updates Patch-set: 3 Topic: interwiki-map
    • Diffusion Commit
    • ·Closed
    Update patch set 3 Patch Set 3: Thank you! Will do. By the way, what are the other tests that you need to do? Patch-set: 3
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: (1 comment) Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 3 Patch-set: 3 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 3 Uploaded patch set 3. Patch-set: 3 Subject: Add mw.site.interwikiMap Commit: c58c528d289d274c09542715c176c697884565bc
    • Diffusion Commit
    • ·Closed
    Create patch set 2 Uploaded patch set 2. Patch-set: 2 Subject: Add mw.site.interwikiMap Commit: 94f0e6fb37e3ee171b331f4097e0b16ce89dfbd5
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: (10 comments) Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: This is the first time I've really done anything in PHP, so please be on the lookout for newbie errors, security problems, code style, and well, everything really. Be ruthless. :) Patch-set: 1 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create change Uploaded patch set 1. Patch-set: 1 Change-id: Ie8ad2582aaf5e422824f7da51714a347bb4041d1 Subject: Add mw.site.interwikiMap Branch: refs/heads/master Commit: 337313ffb1ff330309d3f06016102579332ce9cd
    • Diffusion Commit
    • ·Closed
    Final NoteDb migration updates Patch-set: 2 Topic: interwiki-protorel
    • Diffusion Commit
    • ·Closed
    Update patch set 2 Patch Set 2: No changes in patch set 2 - all I did was update the commit message. The previous commit message was pretty unclear, and confused relative URLs with protocol-relative URLs. Patch-set: 2 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 2 Uploaded patch set 2. Patch-set: 2 Subject: Use protocol-relative URLs in mw.site.interwikiMap Commit: ad34c2bd941c760ffc3639108ee807032a70bffe
    • Diffusion Commit
    • ·Closed
    Create change Uploaded patch set 1. Patch-set: 1 Change-id: If0a169cfd1597e09706fc0e6d5cdcc821ba8d06a Subject: Use protocol-relative URLs in mw.site.interwikiMap Branch: refs/heads/master Commit: 4e3424b3bd41f55140bfdaa4aa600bbf682fc9e3
    • Diffusion Commit
    • ·Closed
    Update patch set 4 Patch Set 4: Jackmcbarn: My first thought is that modules building complex enough HTML to warrant this are probably better off making creative use of mw.html:node. Can you give any examples of modules that would benefit from this change? Patch-set: 4 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Update patch set 2 Patch-set: 2 Label: Code-Review=+1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: Code-Review+1 I couldn't spot any problems when looking through the code, testing out the altered functions didn't throw up any surprises, and the test cases all pass for me. Should be fine to merge, in my opinion. Patch-set: 1 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Code-Review=+1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: -Code-Review Ah, I see. Thanks for clarifying. Patch-set: 1 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Code-Review=0
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: Code-Review-1 Anomie's mw.isSubsting patch was merged back in November 2013, so I'm not sure what this is still doing in the review queue. Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 4 Patch Set 4: The benefit to type-checking is that it's a lot easier to debug modules if the mw libraries generate sensible error messages. Remember that we're writing Scribunto for casual coders who may have little experience in Lua or programming in general; any help we can give them in tracking down bugs has to be a good thing. Type-checking would also make this library consistent with the other mw libraries and with the standard Lua libraries. Providing inconsistent interfaces across different libraries is going to be confusing for module writers. If there's a way we can support things like metatables while still giving sensible error messages then we should do that, but otherwise I'd prefer that we keep things consistent. Patch-set: 4 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Update patch set 2 Patch Set 2: We could always do both - allow either a list of parameters or an array as the first parameter. That's done in mw.message:params, and a few other places as well. What do you think about the change to requiring at least one argument? In mw.math.sum, mw.math.product and a few others I think it makes the most sense, but in others I think outputting NaN is equally valid. If everything else is equal I think we should probably do what the math library does, but I'm not dead set on copying the math library interface if there's a good reason to do it differently. Patch-set: 2
    • Diffusion Commit
    • ·Closed
    Update patch set 2 Patch Set 2: Code-Review-1 This needs some more test cases for the type checking, so marking as -1, but uploading so we can talk about the general direction this is going in. I've switched the interface to use mw.math.funcName( a, b, c, ... ), and implemented type checking with a new validateParameters function. This makes sure that there is at least one value passed to the function, and that each value is a number. This also means that I have removed some of the logic that returned NaNs, etc. Both the parameter checking and the interface are in line with Lua's math library, which we should probably follow here. Patch-set: 2 Label: Code-Review=-1
    • Diffusion Commit
    • ·Closed
    Create patch set 2 Uploaded patch set 2. Patch-set: 2 Subject: Add mw.math library Commit: 63deabfd56b6bf71e4926c9019f095c397ac0c74 Groups: 63deabfd56b6bf71e4926c9019f095c397ac0c74
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: Very nice work in writing this! I've just started testing the Lua parts of this, and I'm liking what I see. One thing I wondered, though - how about making the functions callable as p.funcName( a, b, c, ... ) rather than p.funcName{ a, b, c, ... }? That seems somehow more "Lua-like" to me. Also, how about adding type-checking in using libraryUtil.lua? Patch-set: 1 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Update patch set 2 Patch-set: 2 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Code-Review=0
    • Diffusion Commit
    • ·Closed
    Final NoteDb migration updates Patch-set: 1 Topic: parseQueryString
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Abandoned Patch-set: 1 Status: abandoned
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: Ok, I've finished searching for all module-namespace pages containing the text "parseQueryString" (you can see them at https://en.wikipedia.org/wiki/User:Mr._Stradivarius/parseQueryString), and none of them use the i or j parameters. But yes, despite them not being used on WMF wikis, let's document the parameters rather than removing them. Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: Actually, ignore that last part and the part about "limited use" - I was misreading the code. It might be better just to document the parameters instead, but first I'll analyse the results of my search to see if the parameters are actually being used. Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: I'll make a search across all WMF wikis to see whether it is used anywhere. Given the lack of documentation and the limited use of the parameters - they only affect what portion of the + signs in the query string are converted to spaces - I doubt they are in use anywhere, however. Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: I noticed this when looking at the code for I3ed9529de, which I have just started work on again, but I thought it was best as a separate patch. Patch-set: 1 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Uploaded patch set 1. Patch-set: 1
    • Diffusion Commit
    • ·Closed
    Create patch set 1 Patch-set: 1 Subject: Remove undocumented parseQueryString parameters Commit: 330e5dd6b9f6e2716ddb06e326376207cd80729c
    • Diffusion Commit
    • ·Closed
    Create change Patch-set: 1 Change-id: I11241e0f6be262d199f272381706e03c60dbf5f1 Subject: Remove undocumented parseQueryString parameters Branch: refs/heads/master
    • Diffusion Commit
    • ·Closed
    Final NoteDb migration updates Patch-set: 1 Topic: query-global-fix
    • Diffusion Commit
    • ·Closed
    Create change Uploaded patch set 1. Patch-set: 1 Change-id: Ia28fa10388bfc587d95b522bfa8f3524b4a3ee5f Subject: Fix accidental global in mw.uri.parseQueryString Branch: refs/heads/master Commit: d59d8522902b61990376c8b41e4a7d0463cbd6f7
    • Diffusion Commit
    • ·Closed
    Update patch set 1 Patch Set 1: Code-Review+1 This fixed the problem when I tested it on my development box. Patch-set: 1 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Code-Review=+1
    • Diffusion Commit
    • ·Closed
    Update patch set 14 Patch Set 14: (1 comment) Patch-set: 14
    • Diffusion Commit
    • ·Closed
    Update patch set 14 Patch Set 14: (8 comments) Patch-set: 14 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Update patch set 15 Patch-set: 15 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Update patch set 13 Patch Set 13: Code-Review-1 Now that argument translation has been added to the enwiki version, we should add it here too. Patch-set: 13 Label: Code-Review=-1
    • Diffusion Commit
    • ·Closed
    Update patch set 12 Patch Set 12: Code-Review+1 (2 comments) Looks good to me. I think this is ready to merge now. Patch-set: 12 Label: Code-Review=+1
    • Diffusion Commit
    • ·Closed
    Update patch set 11 Patch Set 11: I've simplified the logic slightly when finding the arg tables, and in doing so I've got rid of the fargs, pargs and luaArgs variables. I split the wrapper-matching code out into its own function as well - this was purely to increase code readability, as the function is only called once. I've also changed the code to create a title object from parent:getTitle(), and to use the mw.title library's equality testing to check whether it is a wrapper template. Seeing as how title equality testing is now inexpensive it seems a shame not to use it. Also, I've tried to de-verbosify my comments. Feel free to trim them down further if desired. Patch-set: 11 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 11 Uploaded patch set 11. Patch-set: 11 Subject: Add getArgs from enwiki's Module:Arguments Commit: 18811f6cc1dbe44e1493dc360c7d1e89819501c5 Groups: 18811f6cc1dbe44e1493dc360c7d1e89819501c5
    • Diffusion Commit
    • ·Closed
    Update patch set 8 Patch-set: 8 Label: Code-Review=-1
    • Diffusion Commit
    • ·Closed
    Update patch set 7 Patch Set 7: Code-Review-1 This implementation suffers from the same bug that I discovered today on enwiki's Module:Arguments. If you explicitly erase an argument, it still shows up when iterating with pairs or ipairs. I've added a couple of test cases to demonstrate, but I'm not sure of the best fix at the moment. Patch-set: 7 Label: Code-Review=-1
    • Diffusion Commit
    • ·Closed
    Create patch set 7 Uploaded patch set 7. Patch-set: 7 Subject: Add getArgs from enwiki's Module:Arguments Commit: 4c1ce8082d1a8ae531cb3c8d2d3be423116cb932 Groups: 4c1ce8082d1a8ae531cb3c8d2d3be423116cb932
    • Diffusion Commit
    • ·Closed
    Create patch set 6 Uploaded patch set 6. Patch-set: 6 Subject: Add getArgs from enwiki's Module:Arguments Commit: eb6e77eda274066f6f75186267959d7a2e9129e6 Groups: eb6e77eda274066f6f75186267959d7a2e9129e6
    • Diffusion Commit
    • ·Closed
    Update patch set 3 Patch Set 3: (3 comments) I mostly agree with the proposed changes to getArgs.lua. Especially, the removal of parent frame access when we don't know what the parent is seems like a good idea. The wrappers option is a superior way of dealing with this, and removing the old behaviour now will allow module writers to update their modules without breaking anything. I do wonder if some module writers would want to have the option to make frame args have precedence over parent frame args when a wrapper template is specified. This is how I originally conceived of enwiki's [[Module:Protection banner]] working, for example, although now it uses a config module for configuration of its various wrapper templates. No plus or minus mark for now, as I've only checked the code visually. I'll report back after I've played around with it a bit, and after I've managed to get phpunit working again. Patch-set: 3 Label: Code-Review=0
    • Diffusion Commit
    • ·Closed
    Final NoteDb migration updates Patch-set: 7 Topic: T93059
    • Diffusion Commit
    • ·Closed
    Update patch set 7 Patch Set 7: Originally I just wanted mw.uri.new to produce the same output as mw.uri.localUrl et al. I didn't realise that this would mean making breaking changes, though. The really obviously wrong things in wfCgiToArray should be fixed, of course - that shouldn't break any existing uses. I'm wondering what to do about false values, though. For example: mw.logObject( mw.uri.parseQueryString( 'foo' ) ) --> { foo = false } var_dump( wfCgiToArray( 'foo' ) ); --> array(1) { ["foo"]=> string(0) "" } If we make Lua match PHP then we risk breaking existing modules, and likewise if we make PHP match Lua we risk breaking things in Core and extensions. I suppose the least problematic would be to make Lua match PHP and then search through existing modules (perhaps with a script) to find any affected modules. But then consciously making the two not match would not be the end of the world. Patch-set: 7
    • Diffusion Commit
    • ·Closed
    Update patch set 7 Patch Set 7: So, should the goal here be to mirror wfCgiToArray precisely? I didn't do that this time because it would contradict with the existing documentation.[1] In particular: "Keys in the string without values will have a value of false; keys repeated multiple times will have sequence tables as values". wfCgiToArray maps keys without values to the empty string, and if keys are repeated multiple times ("foo=bar&foo=baz") then only the final one is used. Some people might already rely on that behaviour, although I don't personally know anywhere where it is used. [1] https://www.mediawiki.org/wiki/Extension:Scribunto/Lua_reference_manual#mw.uri.parseQueryString Patch-set: 7 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 7 Uploaded patch set 7. Patch-set: 7 Subject: Standardise URI query formatting Commit: aba5e7ad6f84452b5a331d598843be73178a9325 Groups: aba5e7ad6f84452b5a331d598843be73178a9325
    • Diffusion Commit
    • ·Closed
    Update patch set 5 Patch Set 5: (1 comment) Patch-set: 5
    • Diffusion Commit
    • ·Closed
    Update patch set 5 Patch Set 5: (3 comments) Patch-set: 5 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Update patch set 6 Patch-set: 6 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 6 Uploaded patch set 6. Patch-set: 6 Subject: Standardise URI query formatting Commit: 7bdc53c58c8326f6a45e5de8c71d1861d8de5266 Groups: 7bdc53c58c8326f6a45e5de8c71d1861d8de5266
    • Diffusion Commit
    • ·Closed
    Create patch set 5 Uploaded patch set 5. Patch-set: 5 Subject: Standardise URI query formatting Commit: 506dd4e8e9366c51c14af0c0646dbfbc2e8eeba1 Groups: 506dd4e8e9366c51c14af0c0646dbfbc2e8eeba1
    • Diffusion Commit
    • ·Closed
    Create patch set 4 Uploaded patch set 4. Patch-set: 4 Subject: Standardise URI query formatting Commit: 47f27ebf6672d6d3bb8f6e867fffa9746025c632 Groups: 47f27ebf6672d6d3bb8f6e867fffa9746025c632
    • Diffusion Commit
    • ·Closed
    Update patch set 3 Patch Set 3: (1 comment) Patch-set: 3
    • Diffusion Commit
    • ·Closed
    Update patch set 3 Patch Set 3: Code-Review-1 (1 comment) Patch-set: 3 Label: Code-Review=-1
    • Diffusion Commit
    • ·Closed
    Update patch set 3 Patch Set 3: In the end, for the second example above, I ended up making a separate sub-table to hold both values. That seemed the logical way to do it based on the existing code. Patch-set: 3 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 3 Uploaded patch set 3. Patch-set: 3 Subject: Standardise URI query formatting Commit: f42b20ed826af1006c9a2eac16e8371bf0b52625 Groups: f42b20ed826af1006c9a2eac16e8371bf0b52625
    • Diffusion Commit
    • ·Closed
    Update patch set 2 Patch Set 2: There are some ambiguous cases that a parseQueryString will have to deal with if we update it to parse parameters with square brackets. Namely: ?foo[0]=bar&foo=baz ?foo=bar&foo[0]=baz In the first one, we have the choice of overwriting the first value, adding a new index to the query table, or raising an error. In the second one, we can either overwrite the first value or raise an error. I'd go with adding a new index to the query table for the first one, and overwriting the value for the second. But I'd like to hear others' opinions before coding this up. Patch-set: 2
    • Diffusion Commit
    • ·Closed
    Update patch set 2 Patch Set 2: (1 comment) Patch-set: 2 Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Create patch set 2 Uploaded patch set 2. Patch-set: 2 Subject: Standardise URI query fields with multiple values Commit: fd8a7e8a295a9c2fc68e172245e640ff890a816b Groups: fd8a7e8a295a9c2fc68e172245e640ff890a816b
    • Diffusion Commit
    • ·Closed
    Create change Uploaded patch set 1. Patch-set: 1 Change-id: I3ed9529deb7a63dbabdc58d7afb3b0cea9307a7f Subject: Standardise URI query fields with multiple values Branch: refs/heads/master Commit: 25b9301a7fef8dfa2476db3db8b237af74c6d5ea Groups: 25b9301a7fef8dfa2476db3db8b237af74c6d5ea
    • Diffusion Commit
    • ·Closed
    Update patch set 14 Patch-set: 14 Label: Code-Review=0
    • Diffusion Commit
    • ·Closed
    Update patch set 13 Patch-set: 13 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Code-Review=0
    • Diffusion Commit
    • ·Closed
    Update patch set 2 Patch Set 2: (1 comment) Patch-set: 2 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Verified=0
    • Diffusion Commit
    • ·Closed
    Update patch set 11 Patch-set: 11 Reviewer: Mr. Stradivarius <1590@e9e9afe9-4712-486d-8885-f54b72dd1951> Label: Code-Review=0
    • Diffusion Commit
    • ·Closed
    Add a vagrant role for the SandboxLink extension Change-Id: I276e5e133033c6cacaf5dabff71c1184ce5d02bf
    • Diffusion Commit
    Follow sandbox link redirects Some users set up redirects at the default sandbox location that point to a custom sandbox location. This change allows them to reach their custom sandbox with one click instead of being shown the redirect page. Bug: T99335 Change-Id: I8ca2ab43d14afbbc2b97429a999cb4f3627676ce
    • Diffusion Commit
    • ·Closed
    Use the new registration format Convert PageBlock.php to extension.json as detailed at https://www.mediawiki.org/wiki/Manual:Extension_registration Change-Id: I97933aea8a73b33ad8b0cc18d1ec27e85f5213e5