Describe the bug
loaded an object file with batch loader from an ar archive. ARM:EB:32
decompiler complains
deleting and reloading didnt help
Screenshots

Attachments
ARM:BE:32:v8 (1.103)
ARM mode
address : [000105c8 - 000106d0]
bytes:
e9 2d 40 70 e5 9f 50 fc e1 a0 40 00 e5 95 36 1c e1 50 00 03 23 a0 40 00 3a 00 00 01 e1 a0 00 04 e8 bd 80 70 eb 00 06 8c e5 d5 36 40 e3 53 00 00 1a 00 00 21 e2 83 30 01 e5 c5 36 40 eb 00 06 87 e5 95 36 14 e7 93 41 04 e3 54 00 00 0a 00 00 0e eb 00 06 81 e5 d4 30 d4 e3 53 00 00 0a 00 00 1c eb 00 06 7e e2 84 00 d0 e3 a0 10 01 eb 00 06 73 e1 d4 31 b8 e2 13 60 01 0a 00 00 1e e3 a0 2e 15 e1 94 30 b2 e2 83 30 01 e1 84 30 b2 eb 00 06 72 e3 a0 3d 19 e1 95 20 b3 e5 9f 50 68 e3 52 0c 01 0a 00 00 0f eb 00 06 6d e2 85 0e 63 e2 80 00 0c eb 00 06 64 e1 a0 00 04 e8 bd 80 70 eb 00 06 67 e2 85 0e 63 e3 a0 10 01 e2 80 00 0c eb 00 06 5b ea ff ff da e2 83 30 01 e5 c4 30 d4 eb 00 06 5f ea ff ff e2 e2 43 3d 19 e5 c5 36 40 eb 00 06 5b e1 a0 00 04 e8 bd 80 70 e2 84 00 d0 eb 00 06 51 e1 a0 40 06 ea ff ff e0 00 01 20 00
Environment (please complete the following information):
cannot repro with the provided bytes/address using f7f366b261589a65696c17852812566b9e6659bb
maybe your .sla file got corrupted, I would delete Ghidra/Processors/ARM/data/language/ARM8_be.sla, restart ghidra, and see if the issue persists
interesting . is there some others caching at project levels ?
$ md5sum ARM8_be.sla
4842592248245a8637a666eaf34b33ef ARM8_be.sla
closed ghidra & del sla
rm ARM8_be.sla
run ghidra && load project & file & new sla get created:
$ md5sum ARM8_be.sla
4842592248245a8637a666eaf34b33ef ARM8_be.sla
here the
exported program as 'Ghidra Zip File' then gzip it :
trsockex.o.gzf.gz
had the same issue occurring on others files from very widely distributed production grade library
@pabx06 Could you also try this using the official 9.1.2 distro from ghidra-sre.org?
I'm getting this as well in 7357b755f4ac658ac3e630d301160f16070ff842. It is difficult to reproduce but it is from a change sometime last week. It does repeatedly occur on the same function so I can reproduce for debugging but unfortunately cannot distribute. It is occurring on x86:LE:64:default (2.9) but I don't think it is sleigh related.
I think it may be related to some of the references which would be why the bytes for the function didn't reproduce it.
This is an example of an XmlElement in question XmlElementImpl@0x72 "<high(3) repref="0x171a" class="global"> @(19478:77008)"
I'm having a hard time tracking down what it would be. None of the files which reference repref or symref have been changed in months :/
i starts to wonder if it is caused by Analyzers not playing nicely with each other on per settings base
I'm reproducing this.
This is fixed now in master. Thank you for the submitting this.