Gene ID in cuffmerge

[Karli_40]

Hi everyone
I have a question on IDs in cuffmerge output. I am doing cufflinks with -g annotation.gtf file. gene_id and transcript_id seem correspond to the gtf file.

2L Cufflinks exon 7529 8116 1 + . gene_id "FBgn0031208"; transcript_id "FBtr0300689"; exon_number "1"; FPKM "0.0000000000"; frac "0.000000"; conf_lo "0.000000"; conf_hi "0.001241"; cov "0.000000";
2L Cufflinks exon 8193 9484 1 + . gene_id "FBgn0031208"; transcript_id "FBtr0300689"; exon_number "2"; FPKM "0.0000000000"; frac "0.000000"; conf_lo "0.000000"; conf_hi "0.001241"; cov "0.000000";

But to run cuffdiff I run cuffmerge first. Then the ids are changed to cufflinks id, which looks like (although I still used the -g option):
2L Cufflinks exon 7529 8116 . + . gene_id "XLOC_000001"; transcript_id "TCONS_00000003"; exon_number "1"; gene_name "CG11023"; oId "FBtr0300690"; nearest_ref "FBtr0300690"; class_code "="; tss_id "TSS1"; p_id "P1";
2L Cufflinks exon 8193 8589 . + . gene_id "XLOC_000001"; transcript_id "TCONS_00000003"; exon_number "2"; gene_name "CG11023"; oId "FBtr0300690"; nearest_ref "FBtr0300690"; class_code "="; tss_id "TSS1"; p_id "P1";

As result when I used this merged gtf file to run cuffdiff, ids still remain in this format...

I am running everything with annotation gtf file when -g option is available so I am not sure why this is still happening.
Thanks for help in advance.