"E36 Not enough room" while sourcing handcrafted session file

[Tony Mechelynck]

Since a few days, while sourcing the following session file:

" Vim session file
" This one was not generated by :mksession.

" e ~/.mozilla/seamonkey/eiox68r9.default/user.js
e ~/.mozilla/seamonkey/r0hh40it.frg-2018+/user.js
lcd %:h
new chrome/userChrome.css
lcd %:h
new userContent.css
new
setl buftype=nofile
new ~/pub/chroniq/froiss/F4_021.htm
lcd %:h
new ~/Session.vim
lcd ~
new ~/calendrier.2024.1112.txt
lcd ~
bot h
if exists('+switchbuf')
set swb=useopen
endif
wincmd W

I get error "E36 Not enough room" at line 7, but the error is not
visible because the GUI opens with just enough width and height for
the menu. If I don't maximize the window manually I get it again at
line 9 etc. (i.e. every ":new" command with file name). My vimrc
includes (among others) an autocommand "au GUIEnter * set lines=99
colums=999" in order to start the GUI maximized. This is a Huge build
with GTK3 GUI and no interpreted languages (i.e. no perl, no python,
no mzscheme, etc.).

I didn't report this problem immediately (I should have) but it is not
long ago (a few days) that it started appearing.

Best regards,
Tony.

[Tony Mechelynck]

P.S. The problem disappears if I add the following line between lines
6 and 7 above:

set lines=99 columns=999 " to work around a GUI bug

and gvim then loads again the files which were omitted because of the
E36 "errors".

Best regards,
Tony.

[Christian Brabandt]

Can you respond to the mentioned ticket? How bad is this? Is this
something we need to consider to revert?

Thanks,
Christian
--
QOTD:
"What women and psychologists call `dropping your armor', we call
"baring your neck."

[Tony Mechelynck]

My problem is not the same as the ticket AFAICT. See my latest message
in toothpik's "GTK error" thread. When I start gvim with -u NONE or -u
NORC or -u DEFAULTS there is no error message on stderr but the
problem (a GUI of minimum size) happens anyway.

My GTK3 version is 3.24.49+14-1.1 and I'm running on openSUSE Tumbleweed.

I think this is bad (though I have found a workaround for my usecase)
and I would have expected other users (not only toothpik and me) to be
hit by it.

Best regards,
Tony.

[Tony Mechelynck]

At patchlevel 9.1.1378 the problem has disappeared, even without the
workaround in my session file.

Best regards,
Tony.