* On 05.07.2014 01:59 am, Mihai Moldovan wrote: > * On 05.07.2014 01:35 am, Karl Berry wrote: >> If you comment out the call to run_recode(), does it solve your problem? >> (Please don't educate me about how this doesn't deal with the "real" >> issue. :) > I'll check that. There may be more invocations of sed failing, so I can not > give > a clear answer without testing. Will reply later.
Not executing run_recode() seems to work fine with the input file/for the failure I experienced. In general, removing run_recode() and friends seems to fix the issue just fine, other code running sed on input files are, is far as I could see, not parsing the texi input file but other intermediate files. Despite of this, I still believe making sed configurable at build time would help counter such issues in a nice and clean way. I'm just thinking about new code added in the future. Mihai
smime.p7s
Description: S/MIME Cryptographic Signature
