[tex-k] segfault in texlive2020 pdftex/luatex/xetex/ptex/...

David Carlisle d.p.carlisle at gmail.com
Mon Jun 1 09:09:41 CEST 2020

On Mon, 1 Jun 2020 at 02:42, Ken Brown <kbrow1i at gmail.com> wrote:

> I'm happy to do whatever Cygwin users want if the default stack size seems
> too
> small.
> Ken

I think I'd only ask for an increased stack if there were documents that
are working on other builds but failing on cygwin. I've used the cygwin tex
for a long time and never seen that.  If you are using this much stack you
are almost certainly looping so having a bigger stack just means it takes
longer to error. In the real document where I first noticed this I passed
#1 instead of ##1 into a nested definition which caused a loop that just
happened to hit this, somehow managing to miss the usual tex limits such as
input or parameter stacks.

The only thing I'm a bit unclear about is whether there is any usable
relationship between the stack size as compiled in to the executable and
the the expand_depth setting or whether it depends on too much other stuff
going on. Also is it documented anywhere? (I don't see in the web2c manual)
That is, is it possible to default expand_depth to a value that makes the
stack overflow less likely based on the compile time values used rather
than experimentally setting it per machine. But as it just changes the
error message and prevents the stackdump file I doubt most users will mind
much either way.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://tug.org/pipermail/tex-k/attachments/20200601/aa760f34/attachment.html>

More information about the tex-k mailing list.