search

gap-system / gap

Main development repository for GAP - Groups, Algorithms, Programming, a System for Computational Discrete Algebra
https://www.gap-system.org
GNU General Public License v2.0
763 stars 159 forks source link

support initial search in `PAGER_BUILTIN` #5683

Closed ThomasBreuer closed 1 month ago

ThomasBreuer commented 2 months ago

When PageSource calls Pager with the data for a kernel function, it specifies the start line by a search string. This works well for external pagers such as less, but up to now PAGER_BUILTIN did not honor this hint.

With the changes proposed here, calls such as PageSource( IS_SSORT_LIST_DEFAULT ) will work as expected also when PAGER_BUILTIN is used.

(The code is a hack, because GAP does not support regular expressions and because the search in C code files is based only on some heuristics.)

ThomasBreuer commented 2 months ago

@frankluebeck suggested that it is better to support an initial search in general for the builtin pager, but to drop the support for a regular expression. I will update the pull request.

frankluebeck commented 2 months ago

Sorry, I communicated my review in person. I should have left a short note here as well.

frankluebeck commented 2 months ago

So far we are using less with +[initial command] syntax. The initial command can be anything, e.g., a search like /xyz or going to the end G or ... or going to line 13 13 or 13g.

ThomasBreuer commented 2 months ago

The test codecov/project is now running for more than four days. (Started 2024-03-28 17:29:05.) Apparently something went wrong.

ThomasBreuer commented 1 month ago

The test codecov/project is now running for more than 12 days. How can this be stopped/restarted?

fingolfin commented 1 month ago

That the codecov test is reported as "running" doesn't mean something is running somewhere, just that received an event "test has started", but somehow the matching "test has finished" was not delivered (can always happen due to e.g. network issues).

It is also perfectly fine to ignore that test. Normally I would have, but in this case I decided to close&reopen to re-run CI anyway, because something else was weird (all codecov/... checks were listed twice). Strangely enough, this remains, but not for other PRs. Very odd.