There are some hidden characters inside like "়া" (without the last one) which are counted with len() but I have 5 characters available until the rigtht side of a 80 column page. Maybe this is a such a minor problem that we just could decline it. I guess it does not worth to fix it.

We have moved to git and are no longer on svn. Can you show us your output of python pwb.oy version?

py 3.3 is no longer supported by pwb

See comments above

Decoding sys.argv[0] would solve this issue above
absolute_path = abspath(os.path.dirname(sys.argv[0].decode('utf-8')))
but the same problem occurs when importing:

This is a core 2.0 issue which is no longer maintained.

In T106121#4445145, @Dvorapa wrote:

Okay, now we can remove deprecations older than 2 ( ? 3? ) years!

Could you explain a bit please.

Is there a reason to leave her?

I guess the manual at mediawiki is a better place for the manual than docs because it is easier to update and it sin not needed to submit changes throught the code review workflow. Whereas docs is the right place for the api documentation and additional stuff related to it. I tend to remove that warning banner.

See also T130911

I've added the L10N for content dict to use clean_sandbox for "bn" site code. This works for all projects wich have a sandbox page listed on wikidata item Q3938. For he others the -page option has to be used to choose the page. Or create a wikidata item for it which could be added to the "sandbox_titles" item tuple.

-easy: this is not trivial. First we need the common content for the sandbox page but what is it? Either one for all or one for each site or one for each page. Currently 'bn' isn't localized. Second the default project page is either Q3938 or Q28939665. Now we have 3 pages for each 'bn' site. Finally multipe -page options may work but the content is still missing and -text option is needed.

In T134172#4442802, @Dalba wrote:

I guess 2be2bbda8de16408723074a4ebcdf1d85e403fb0 has resolved it.

This is valid for compat only

I'm confused, there already are some deprecations in 7 scripts. So we just need to start deprecating in scripts? I would support to finish https://gerrit.wikimedia.org/r/#/c/pywikibot/core/+/399755/ ASAP for this.

In T57212#4429962, @Dvorapa wrote:

At first this needs to add localisation into commonscat.

There are two L10N dicts in commonscat already. What is needed further?

Maybe we should move scripts directory to _scripts to make this more explicit?

In T106121#4430075, @Dvorapa wrote:

Personally I would continue to deprecate parts of scripts as before.

Can somebody update what is to be updated here please.

I also cannot follow this request. Could you explain please or make an example of what do you expect and what is the current behavior.

Is that implementable?

bot and redirect parameters are already implemented for newpages. Adding this to NewpagesPageGenerator does not make any sense because GeneratorFactory.handleArg cannot retrieve these parameters.

In T198889#4426746, @MarcoAurelio wrote:

could we have some instructions as to how to configure those projects on the clean_sandbox.py script code? Thanks.

See that sample given above:

In T198889#4426025, @Xqt wrote:

It can have different settings already. Just add it to the xdict L10N table „content“ like:

Ah, I see: The problem is the check inside initializer.

Cannot follow this task. Pywikibot is throttleling not to edit more than 6 (I guess) pages per minute by default. What might be the reason that the limit of 40 pages is exceeded. What was the putthrottle setting when the exception occurred?

It can have different settings already. Just add it to the xdict L10N table „content“ like:

I did some measurements for the memory needed for the statements given above:

We could

• use hash function for the filter_unique key
• use hash function for the filter_unique key by default
• use a GeneratorFactory Container attribute to hold the seen pages which could be reused when we have more than one duplicate filter
• use an container which uses disk space instead of memory (but this could be time consuming)

No glue where the Memory leakage might come from.

I see the getsizeof() counts the pointers only but not the Page objects itself.

Long-running tasks may end on MemoryError to filter_unique leaks Memory

Why do you assume that?
Try:

This is too undetermined. Could you give a sample?
When filter_unique is used the filter key could be replaced by the hash but this does not solve a lot. Most space is used by the content but it is not clear to me where this is hold after processing.

unsigned comments are valid threads and archivebot ignores them. Unfortunately these threads leads test to fail but we must not ignore them otherwise we cannot check whether the regexes work properly. How proceed in this case?

• either add a timestamp to the talk thread
• or use an alternate talk page for a given site

All other exceptions means there is an undetected timestamp and should be fixed be the regex.

The error message is equal to that in T69663. Can we solve it at once?

https://api.travis-ci.org/v3/job/402542460/log.txt

No glue why this rdirection change was introduced with T129193; I propose to revert it as suggested.

Seems it is solved upstream. No further failure seen in tests.

Seems this edit shouldn't have been done too.