Re: etex and pdftex have diverging memory constraints?
jfbu via tex-live <[email protected]>
| Newsgroups | gmane.comp.tex.live |
|---|---|
| Message-ID | <[email protected]> |
> Le 27 août 2025 à 17:11, jfbu <[email protected]> a écrit : > > > You may notice this differs from what I reported initially. I do not know what happened. I think I did a rebuild of the etex format after my initial post in order to check the impact of main_memory. > Sorry for some last noise : when I wrote the initial message I was in the development repository of xint. Although the macros used in the test file have not been modifed since 2022 and the last upload to CTAN, the file xintkernel.sty itself has some unrelated updates. This impacts the available TeX memory hence the reported maximal value 2492694 for compilation with etex and 2492678 with pdftex. But with the xintkernel.sty from CTAN it is found to be on my system 2492701 with etex and 2492685 with pdftex. And trimming the xintkernel.sty to only contain the needed support for the two used macros (which leaves less. than 90 lines) I get 2494089 as maximal value with etex and 2494074 with pdftex. Ultimately my conundrum is probably resolved by what @JonathanFine said about the format taking out some TeX memory. On my system etex.fmt has 1552667 bytes and pdftex.fmt has 1552830 bytes. This is 167 bytes more but the limit in my test is only at about 15 tokens less allowed as JF